Review of indicators to assess Infant Feeding in Emergencies

Summary of research¹

Population movements in an emergency challenge needs assessments - including infant and young child feeding practices © ANDY SEAL

The public health importance of infant feeding in emergencies has been highlighted in countries such as Iraq and Bosnia where feeding infants with breastmilk substitute is common practice. Although there are few epidemiological studies on the impact of emergencies on infant feeding, many anecdotal reports of adverse health outcomes exist. The displacement of such populations has created new dilemmas for aid workers on how best to assess and support feeding practice.

During the 1999 Kosovo Crisis an opportunity arose to research humanitarian interventions in infant feeding in the Former Yugoslav Republic (FYR) of Macedonia.

Above and below: standard feeding indicators do not adequately assess artificial feeding practices of a population. Refugee camp, FYR Macedonia, 1999

Indicators, recommended by international health and nutrition organisations for assessing infant feeding practice, were compiled and analysed to evaluate their consistency and applicability for use in surveys of emergency-affected populations. These included measures of breast-feeding status, use of artificial feeding, anthropometric status and morbidity. In addition, health and nutrition status surveys performed on the resident or refugee population of Kosovo during the years 1996-1999 were reviewed to compare their use of infant feeding and morbidity indicators.

Comparison of standard indicators

Indicators recommended by the World Health Organisation (WHO), UNICEF Multiple Indicator Cluster Survey (MICS) Indicators for Global Reporting, Wellstart International Expanded Programme on Breastfeeding (EPB) and Measure Demographic Health Surveys (DHS) were compared. This revealed a number of inconsistencies, both in target population and measurement method.

For example:

• The recommended age-group in which exclusive breastfeeding rate was measured varied between less than four months (WHO, UNICEF), and less than six months (EPB, DHS), thus limiting comparability of data.
• The standard indicator for timely complementary feeding, recommended by both WHO and UNICEF, refers to breastfed infants only. Thus in a population where a proportion of the infants are solely artificially fed, they would be excluded from this assessment.
• There is currently no specific definition of diarrhoea for infants under six months. The WHO standard definition of acute watery diarrhoea (three or more loose stools in 24 hours) closely resembles the minimum number of stools normal for an effectively breastfed infant (three or more stools in 24 hours).² This greatly limits the interpretation of morbidity data in young infants.
• Timely initiation of breastfeeding is not included as a MICS global indicator and definition varies between DHS and WHO, and Wellstarts EPB.

Use of indicators during Kosovo crisis

The use of recommended indicators during the Kosovo crisis was inconsistent, with many nonstandard indicators and methodologies used instead.

For example, a large inter-agency nutrition and health survey, carried out in seven refugee camps in FYR Macedonia in 1999, included no standard indicators on infant feeding, while an infant feeding and weaning survey of the returned population in Kosovo included only one recommended infant feeding indicator. Seven of the reviewed surveys measured exclusive and predominant breastfeeding rates, however only two actually used the 24 hour recall method as recommended by the WHO.

Age-groups for which feeding and morbidity data were gathered also varied widely and infants under six months were often not included in anthropometric or feeding practice assessments. In some cases, feeding practice of young infants was inferred from data gathered from older infants and young children.

These limitations did not prevent comparison, interpretation or conclusion on infant feeding practice in survey reports. Conclusions appeared to be based on field perceptions and experiences rather than on actual data collected.

Discussion

The authors suggest that the inconsistencies observed may reflect a lack of awareness by personnel of current recommendations. The widespread use of non-standard indicators may also reflect gaps in the scope of current assessment tools in emerging emergency situations. Although many standard infant feeding indicators have been developed and are widely used in non-emergency settings (e.g. UNICEF Baby Friendly Hospital Initiative), in reality they have not yet been operationalised in the context of emergencies. Also, current indicators have been developed particularly to assess breastfeeding practice but few recommend how to assess the extent and nature of artificial feeding in a population. The use of a number of non-standard indicators in Kosovo may have reflected a need to assess artificial feeding practice, for which standard indicators do not exist.

Although the benefits of appropriate infant feeding in terms of child survival are well known, the evaluation of aid impact in terms of morbidity and mortality may not be practical or feasible in emergencies. Impact indicators, such as morbidity and anthropometry, have particular constraints when applied to young infants. Furthermore, there may be many indirect influences on infant and child feeding practice in an emergency situation. Monitoring of the entire aid process is necessary to evaluate impact, assign responsibility and encourage accountability. In particular, there is a need to broaden the field concept and practice of evaluation to include process indicators (e.g. number of mothers enrolled in a breastfeeding programme) and outcome indicators (e.g. breastfeeding rates) of infant and child feeding practice.

Recommendations

The authors conclude that during the Kosovo crisis, an inconsistent approach to assessment and monitoring prevented conclusions being drawn about the effectiveness of the international response in protecting infant health and nutrition.

Suggestions to improve future monitoring are made, including:

• Recommended outcome indicators and sampling strategies for assessing infant and child feeding practice should be developed and included in emergency field manuals for health, nutrition, logistics and donor personnel.
• The scope of existing standard indicators is not sufficiently comprehensive to address all feeding issues in emergency situations, particularly in relation to artificially fed populations. Further development and field testing of standard indicators for this population group is necessary if appropriate and comparable assessments are to be made.
• Ultimately, the presence of experienced key personnel in the field is essential to implementing international recommendations and guidelines. This requires significant improvements in field technical support and, where resources permit, early field positioning of an infant and child feeding co-ordinator in emergencies.

¹McGrath M, Seal A, Taylor A (2002). Infant feeding indicators for use in emergencies: an analysis of current recommendations and practice. Public Health Nutrition: 5(3), 365-372

²Evidence-Based Guidelines for Breastfeeding Management during the First Fourteen Days, International Lactation Consultant Management (ILCA), 1999